Dynamical heterogeneous networks

Fraunhofer ITWM

The group dynamical heterogeneous networks is dealing with the modeling and analysis of complex networked systems. This subject (formerly CAD for Analog Circuit Design) has its origin in the field of modeling and analysis of analog circuits. In this context the EDA tool Analog Insydes has been developed. The objective is the integration of symbolic methods into the industrial design flow, thus supporting the circuit designers in their daily work.

The mathematical foundations are mixed symbolic/numerical algorithms for linear and nonlinear differential-algebraic systems of equations (DAE systems), which can also be applied for the modeling and analysis beyond the microelectronics domain. Examples are the automated modeling of regional and national gas pipeline networks, the analysis of complex biochemical reaction networks, as well as electrical power distribution systems and mechatronical systems.

more

There is a large number of various industrial applications of symbolic methods, ranging from classical system understanding to error analysis, and from design and optimization to behavioral modeling on system level. Especially the automatic behavioral modeling is a very promising approach with respect to the simulation of heterogeneous systems (system simulation). This aspect is gaining particular importance for industrial hardware development, because the complex development of hardware prototypes can thus be avoided, including the development costs. A special challenge is heterogeneity, i.e. the coupling of various hardware components from different physical areas. A further research area is the formal verification of digital and hybrid systems which allows for a secure checking of required properties already during the design process.

Projekte

  • VerSys - Verification of Systems-on-Chip with algebraic Methods

    The VerSys project conducts research on formal verification techniques for Systems-on-Chip (SoCs). It innovates on previous approaches by leveraging newly developed methods of computer algebra. The project is focussed on proving functional correctness of highly optimized data paths that comprise the computational heart of today's SoC designs.

  • HIESPANA - Hierarchical simulation of nanoelectronic systems for controlling process variations

    With the transfer from micro- to nanoelectronics the control of production deviations can not keep pace with the reduction of the absolute sizes of semiconductor devices. This results in a drastic increase of relative parameter variation in the system and thus an increasing refuse of produced circuits due to system behaviors beyond the specifications.

  • Analog Insydes

    Analog Insydes is a Mathematica toolbox for modeling, analysis, and design of analog electronic circuits, tailored specifically for industrial applications. Based on a hierarchical netlist description language, you can describe both linear and nonlinear circuits as well as control systems, and automatically set up equation systems in the frequency and time domains.

Competences

  • Symbolic analysis methods
  • Numerical simulation techniques
  • Symbolic approximation of linear and nonlinear DAE systems
  • Automatic generation of behavioral models
  • Hierarchical modeling and simulation techniques
    • Sensitivity analysis
    • Tolerance analysis via interval arithmetic and Monte-Carlo techniques
    • Verification of digital and hybrid systems

    Further Information